Hi, our model is based upon Keras, it is not recommended to be pickled.
Can we save the model with Keras model.save API? Then you can use keras.model.load_model to load it.
Yes please do so and submit we will look at your submission. Please provide clear documentation of how to load your model and put an example of it in your notebook.
Thanks and happy coding ! Lionel.
Yes I’ll document it.
BTW, as for the price competition, is it also required to submit the functions in the qlib payout?
Yes this is required as well. This will improve your explainability score and thus your ranking.
Thanks for your imediate reply.
And should we directly call this function defined in the qlib? So there is no function content in the notebook?
Yes you can define this function in Qlib that will be even better ! Thanks.